Quality Estimation of Virtual Machine Placement in Cloud Infrastructures

Jorge Lopez (1, 2, 3) , Natalia Kushik (4, 2, 3) , Djamal Zeghlache (5, 2, 3)

1 SAMOVAR - Services répartis, Architectures, MOdélisation, Validation, Administration des Réseaux
2 TSP - RS2M - Département Réseaux et Services Multimédia Mobiles
3 CNRS - Centre National de la Recherche Scientifique
4 METHODES-SAMOVAR - Méthodes et modèles pour les réseaux
5 R3S-SAMOVAR - Réseaux, Systèmes, Services, Sécurité

A virtual machine (VM) placement module is a component/part of a cloud (computing) infrastructure, which chooses the best host(s) to allocate the requested VMs. In the literature, skewed or biased criteria are often used to determine the correctness of a placement module. Therefore, the quality of existing placement solutions is not always assessed adequately. In this paper, we propose a distance function that estimates the quality of the placement by comparing it with an optimal solution. We show how this distance function is utilized for testing and monitoring the behavior of VM placement implementations. To validate our approach a simulator has been developed and used for estimating the quality of different placement modules running under various scenarios. Preliminary experimental results on VM placement algorithms implemented in widely used platforms, such as OpenStack show that very often VMs are placed very far from the optimal solutions.

Mots clés: Quality estimation; Distance functions; Integer linear programming; Virtual machine placement; Testing; Monitoring

Domaines: Informatique [cs]; Réseaux et télécommunications [cs.NI]

Jorge Lopez, Natalia Kushik, Djamal Zeghlache. Quality Estimation of Virtual Machine Placement in Cloud Infrastructures. 29th IFIP International Conference on Testing Software and Systems (ICTSS), Oct 2017, St. Petersburg, Russia. pp.213-229, ⟨10.1007/978-3-319-67549-7_13⟩. ⟨hal-01678958⟩